| Quantity: | |
|---|---|
Chassis | Form Factor | 1U Rackmount |
Processor (per Node) | CPU | Single processor(s) |
Note | Supports up to 500W TDP CPUs (Liquid Cooled) | |
System Memory (per Node) | Onboard Memory | |
On-Board Devices (per Node) | NVIDIA C2 | |
Input / Output (per Node) | LAN | 1 RJ45 1 GbE Dedicated BMC LAN port |
USB | 1 port(Rear) | |
Video | 1 mini-DP port | |
System BIOS | BIOS Type | AMI 64MB SPI Flash EEPROM |
Management | Power configurations | Power-on mode for AC power recovery |
PC Health Monitoring | CPU | Monitors for CPU Cores, Chipset Voltages, Memory |
FAN | Fans with tachometer monitoring | |
Temperature | Monitoring for CPU and chassis environment | |
Dimensions and Weight | Height | 43.6 mm |
Width | 438.4 mm | |
Depth | 900 mm | |
Package | 229.87mm (H) x 694.944mm (W) x1169.92 (D) | |
Weight | Gross Weight: 22.32 kg | |
Available Color | Silver | |
Front Panel | Buttons | Power On/Off button |
Expansion Slots (per Node) | PCI-Express (PCIe) Configuration | Default |
Power Supply | 2x 2000W | 2x 2000W Redundant Titanium Level (96%) |
Dimension (WxHxL) | 73.5 x 40 x 185 mm | |
Input | 1000W: 100-127Vac / 50-60Hz | |
Output Type | Backplanes (gold finger) | |
Operating Environment | Environmental Spec. | Operating Temperature: 10°C to 35°C |
Certification | Certifications | CE, FCC, CCC, RoHS |
1U-VGS-121L is a high-density, liquid-cooled 1U rackmount server with 2 independent nodes, each powered by one NVIDIA Grace™ CPU Superchip (72‑core ARMv9). It is designed for extreme performance, efficiency, and scalability in modern data center, AI, and HPC environments.
Form Factor: 1U rackmount, 2 nodes per chassis (ultra‑dense computing)
CPU: 1× NVIDIA Grace CPU Superchip per node (72‑core ARM, up to 3.0 GHz)
Inter‑Node Interconnect: NVLink C2C at 900 GB/s (low latency, high bandwidth)
Memory: Up to 480 GB LPDDR5X (onboard, ECC) per node
Storage:
Up to 4× hot‑swap E1.S NVMe bays per node
2× M.2 NVMe slots per node
Networking/Acceleration:
2× PCIe 5.0 x16 slots per node
Supports NVIDIA BlueField®‑3 DPU or NVIDIA ConnectX®‑7 NIC
Cooling: Direct liquid cooling (DLC) for CPUs (high power efficiency, lower PUE)Supermicro
Power Supply: 2700W redundant PSU (1+1)
Management: IPMI 2.0, Redfish, remote KVM
High‑Performance Computing (HPC)
Scientific computing, computational fluid dynamics (CFD), finite element analysis (FEA)
Weather modeling, quantum chemistry, oil & gas exploration
AI & Generative AI Infrastructure
Large Language Model (LLM) training/inference
Generative AI (AIGC), computer vision, recommendation engines
AI cluster building (MGX‑compliant)
Hyperscale Cloud & Data Analytics
Cloud‑native workloads, big data processing (Spark, Hadoop)
In‑memory databases, real‑time analytics
Enterprise Data Centers & Virtualization
High‑density virtual machines (VMs), containerized workloads (K8s)
Database acceleration, middleware, and cloud gateway services
2× Performance in 1U: Dual nodes + Grace Superchips deliver industry‑leading density
Liquid‑Cooled Efficiency: Lower TCO, reduced noise, and higher reliability
NVIDIA Ecosystem Ready: Native support for BlueField‑3/ConnectX‑7 and NGC software
ARM Architecture: Optimized for modern AI/cloud workloads
